Specialty Courts Recidivism Study: 2014-2021 Cohorts

Author

Adam S. Cohen

Published

May 15, 2025


Abstract

This study investigates patterns of recidivism and probation success within three of the First Circuit Specialty Courts: Hawaii Drug Court (HDC), Mental Health Court (MHC), and Veterans Treatment Court (VTC). Generally, recidivism rates did not change over the study period within the Specialty Courts. Although the recidivism rate sat at 52.3%, 43.8%, and 33.3% for HDC, MHC, and VTC, respectively, only 12.2%, 5.6%, and 7.7% of clients were convicted of new charges or had their probation revoked. Time to recidivate averaged between 10.9 and 12.1 months. Property offenders were more likely to recidivate than drug or “other” offenders, but no other differences were statistically detectable. There were no sex/gender differences in recidivism, although there were a few race/ethnicity and age differences. Among graduates of the Specialty Courts, recidivism rates dropped to 36.4%, 27.9%, and 22.6% for HDC, MHC, and VTC, respectively. Limitations in comparing these data to historical statewide recidivism data or future recidivism data, or comparing between the Specialty Courts, are discussed, as well as future directions for conducting rigorous studies of the Specialty Courts.

Introduction

Hawaii’s Specialty Courts provide intensive supervision and support to individuals who have not had success on regular probation. The Judiciary has undertaken an initial study using historical data to establish baseline numbers on recidivism and to investigate the effectiveness of these courts. This study investigates patterns of recidivism and probation success with three of the First Circuit Specialty Courts: Hawaii Drug Court (HDC), Mental Health Court (MHC), and Veterans Treatment Court (VTC).

This report examines recidivism in terms of rearrests, probation revocations, and criminal contempt of court. The sample included probationers admitted to a Specialty Court in the First Circuit between 2014 - 2021. Probationers were monitored over a three-year follow-up period from the date they entered a Specialty Court.

Method

Sample

Recidivism among felony probationers in the First Circuit Specialty Courts was examined over a precise 36-month period starting from the date a probationer was admitted to a Specialty Court. The entire cohort was comprised of 551 probationers from the following Specialty Courts:

  1. 384 HDC probationers
  2. 89 MHC probationers
  3. 78 VTC probationers

Secondary Data Sources

The study data set originated from multiple information systems.

Rearrest data came from the Criminal Justice Information System (CJIS) database, housed with the Office of the Attorney General. Data was pulled in March 2025, and included the listing of any and all state charges reported since each offender’s respective date of admission to the Specialty Courts, as well as the date of arrest and the disposition of charges. The date of the arrest was used to determine the time to recidivate, or length of time elapsed from the follow-up date to the arrest date, and the disposition of the new charge was used to determine the final outcome of the alleged offense. In situations involving multiple charges filed on the same arrest date, the most severe charge (i.e., class A felony, class B felony, or class C felony) became the recorded recidivism event.

Probation data, including a list of probationers along with the Specialty Court they were admitted to, the date of admission, and graduation status, were provided by the Specialty Court Administrator in stand alone spreadsheets. Additional information about supervision cases (e.g., initial offense, severity of charge, demographics, etc.) was extracted from the CX database, which underlies Hawaii State Judiciary’s probation system, in April 2025.

Risk classification (LSI-R) data was pulled from the Community Corrections Adult Assessment information system created by Cyzap Inc. The LSI-R data was pulled in March 2025, and includes risk assessments conducted on Specialty Court probationers going back to 2014, the start of the study period.

Results

Recidivism for the first fourteen analyses was defined as a new arrest, probation revocation, or criminal contempt of court. New arrests were determined by taking the most recent arrest after admission to the Specialty Courts.

Extreme caution is warranted in drawing conclusions about the relative effectiveness of the different courts from this data. Except when statistical tests are reported, these numbers should be treated as basic baseline estimates of recividism, and attempts to compare are likely invalid and should generally be avoided. See the discussion for a detailed explanation of the perils and pitfalls in comparing the courts to each other, to general probation, or to future recidivism rates.

Full Sample

Results for the full sample of clients admitted to the Specialty Courts between CY 2014 - 2021 are reported first. Results for only those probationers who graduated from the Specialty Courts are reported after the results for the full sample.

Analysis 1: Overall Recidivism Rates

Recidivism rates were computed for each Specialty Court between CY2014-2021.

Analysis 2: Recidivism Rates by Yearly Cohort

Recidivism rates for each Specialty Court were disaggregated by cohort, in which a cohort was defined as the calendar year in which probationers were admitted to each Specialty Court. In other words, a probationer admitted into Hawaii Drug Court on September 1st, 2015 would be counted toward the 2015 HDC cohort whereas a probationer admitted into the Mental Health Court on March 18th, 2019 would be counted toward the 2019 MHC cohort.

There were no statistically detectable differences in recidivism rates in HDC between 2014 and 2021. There were also no statistical differences in MHC or VTC, and even when all probationers were combined, there were no differences across cohorts.

The difference in recidivism rates between cohorts for Drug Court was not statistically significant, \(\chi^2\)(7, N = 384) = 8.76, p = 0.271.
The difference in recidivism rates between cohorts for Mental Health Court was not statistically significant, \(\chi^2\)(7, N = 89) = 6.32, p = 0.503.
The difference in recidivism rates between cohorts for Veterans Treatment Court was not statistically significant, \(\chi^2\)(7, N = 78) = 6.68, p = 0.463.
The difference in overall recidivism rates between cohorts was not statistically significant, \(\chi^2\)(7, N = 551) = 12.43, p = 0.0873.

Analysis 3: Time to Recidivism

Recidivism was analyzed by time to reoffense. Longer times to reoffend can indicate gradual desistance from crime.

Analysis 4: Recidivism Rates by Disposition Status

Recidivism rates were analyzed by disposition status. A rearrest does not necessarily indicate new charges, let alone a new conviction. (Of course, failure to pick up a new arrest does not mean a person has not reoffended.) About 40% of new arrests/offenses resulted in acquittal, dismissal, or charges not being filed. Among Specialty Court clients, 12.2%, 5.6%, and 7.7% of them were convicted of new charges or had their probation revoked in HDC, MHC, and VTC, respectively.

All dispositions that appeared in the rearrest records provided by CJIS are listed below. The following mappings from dispositions (values in the columns) to disposition categories (column heads: “Acquitted/Dismissal,” “Pending - Early pretrial,” Pending - Late pretrial,” “Guilty,” or “Other”) were used:

 

Acquitted/Dismissal Pending - Early pretrial Pending - Late pretrial Guilty Other
NO ACTION WAS TAKEN BY THE COURT SUBJECT WAS TAKEN TO THE INTAKE SERVICE CENTER THE CHARGE WAS CONTINUED UNTIL PROBATION REVOKED; SUBJECT WAS RE-SENTENCED SUBJECT WAS RELEASED TO THE CIRCUIT COURT CELLBLOCK
NO ACTION; CHARGE CLOSED BY PROSECUTOR BAIL WAS RESET FOR DRUG COURT CASE SUBJECT WAS RELEASED PENDING FURTHER INVESTIGATION SUBJECT WAS FOUND GUILTY REVOCATION OF BAIL OR SUPERVISED RELEASE WAS DISMISSED
SUBJECT WAS RELEASED; PROSECUTION DECLINED SUBJECT WAS TAKEN TO THE INTAKE SERVICE CENTER SUBJECT WAS RELEASED ON BAIL PROBATION REVOKED; THE ORIGINAL SENTENCE WAS RESUMED OR MODIFIED THE CASE-IN-CHIEF IS COMPLETE BUT NO DISPOSITION WAS FOUND FOR THIS ARREST
THE CHARGE/CASE WAS STRICKEN SUBJECT WAS TAKEN TO DISTRICT COURT FOR ARRAIGNMENT PAROLE WAS REVOKED RELEASE ON BAIL OR SUPERVISED RELEASE WAS MODIFIED OR REVOKED
THE CHARGE WAS DISMISSED/NPQ WITHOUT PREJUDICE THE CASE-IN-CHIEF IS COMPLETE BUT NO DISPO WAS FOUND FOR THIS CHARGE
THE CHARGE WAS DISMISSED A PENAL SUMMONS WAS ISSUED
THE CHARGE WAS NOLLE PROSEQUI
SUBJECT WAS ACQUITTED
SUBJECT WAS RELEASED; NO CHARGE
PROSECUTION WAS DECLINED

 

Analysis 5: Recidivism Rates by Recidivism Type

Recidivism rates were analyzed by recidivism type. Reoffending included new arrests, probation revocations, and criminal contempt of court.

Recidivism rates differed statistically between the specialty courts for some of the recidivism types. Specifically, HDC had higher rearrest but lower revocation rates than the other two courts.

There was a statistically significant difference in criminal arrests between Specialty Courts, \(\chi^2\)(2, N = 551) = 17.39, p < .001.
There was a statistically significant difference in probation revocations between Specialty Courts, \(\chi^2\)(2, N = 551) = 8.77, p = 0.0125.
The difference in criminal contempt of court between Specialty Courts was not statistically significant, \(\chi^2\)(2, N = 551) = 4.65, p = 0.0976.

Analysis 6: Average Time to Recidivate by Recidivism Type

Time to recidivate did not statistically differ between the specialty courts for any of the recidivism types. Specifically, time to rearrest was no different for HDC, MHC, OR VTC. The same was true for revocations and criminal contempt of court. Note that an increase in time to rearrest can indicate gradual desistance from crime.

Time to recidivate was measured in months.

Criminal Arrest: The main effect of specialty_court is statistically not significant and small (F(2, 145) = 1.29, p = 0.278; Eta2 = 0.02, 95% CI [0.00, 1.00]).
Probation Revocations: The main effect of specialty_court is statistically not significant and small (F(2, 44) = 0.39, p = 0.682; Eta2 = 0.02, 95% CI [0.00, 1.00]).
Criminal Contempt of Court: The main effect of specialty_court is statistically not significant and small (F(2, 68) = 0.44, p = 0.647; Eta2 = 0.01, 95% CI [0.00, 1.00]).

Analysis 7: Recidivism Rates by Initial Offense Type

Recidivism rates statistically differed between initial offense categories. Specifically, the recidivism rate was statistically greater for property offenders than drug offenders or “other” felony offenders. All other comparisions were not statistically different. Initial offenses were categorized by chapters and sections of the Hawaii Revised Statutes according to the category mappings tab below.

There was a statistically significant difference in recidivism rates between initial offense types, \(\chi^2\)(3, N = 551) = 18.54, p < .001.
Property vs Drug: \(\chi^2\)(1) = 15.85, p < .001.
Property vs Other: \(\chi^2\)(1) = 5, p = 0.0253.

Analysis 8: Recidivism Rates by Initial Offense Type and Recidivism Type

Recidivism rates did not statistically differ between initial offense categories when broken down by recidivism type. Initial offenses were categorized according to the category mappings described above.

The difference in criminal arrests between initial offense types was not statistically significant, \(\chi^2\)(3, N = 551) = 0.92, p = 0.821.
The difference in probation revocations between initial offense types was not statistically significant, \(\chi^2\)(3, N = 551) = 2.18, p = 0.536.
The difference in criminal contempt of court between initial offense types was not statistically significant, \(\chi^2\)(3, N = 551) = 4.08, p = 0.253.

Analysis 9: Recidivism Rates by Sex/Gender

Recidivism rates did not statistically differ between sex/gender groups in any of the Specialty Courts. Data was missing for two probationers, who were categorized as “UNKNOWN.”

The difference in recidivism rates between male and female Drug Court clients was not statistically significant, \(\chi^2\)(2, N = 384) = 4.38, p = 0.112.
The difference in recidivism rates between male and female Mental Health Court clients was not statistically significant, \(\chi^2\)(1, N = 89) = 0.24, p = 0.627.
The difference in recidivism rates between male and female Veterans Treatment Court clients was not statistically significant, \(\chi^2\)(1, N = 78) = 0, p = 1.
The difference in recidivism rates between all males and all female was not statistically significant, \(\chi^2\)(2, N = 551) = 4.55, p = 0.103.

Analysis 10: Recidivism Rates by Race/Ethnicity

Recidivism rates differed between race/ethnicity groups, but only in Drug Court. Specifically, probationers of all races/ethnicities except African-Americans had lower recidivism rates than Samoans among Drug Court probationers. All other comparisons were not statistically significant.

There was a statistically significant difference in recidivism rates between races/ethnicities among Drug Court clients, \(\chi^2\)(6, N = 384) = 12.64, p = 0.0491.
The following groups in HDC statistically differed:
HDC, Hawaiian vs Samoan: \(\chi^2\)(1) = 7.64, p = 0.0057.
HDC, Caucasian vs Samoan: \(\chi^2\)(1) = 6.51, p = 0.0107.
HDC, Filipino vs Samoan: \(\chi^2\)(1) = 10.06, p = 0.00152.
HDC, Japanese vs Samoan: \(\chi^2\)(1) = 7.41, p = 0.0065.
HDC, All others vs Samoan: \(\chi^2\)(1) = 8.53, p = 0.00349.
All other comparisons between groups in HDC were not statistically different.

The difference in recidivism rates between races/ethnicities among Mental Health Court clients was not statistically significant, \(\chi^2\)(6, N = 89) = 3.42, p = 0.754.
The difference in recidivism rates between races/ethnicities Veterans Treatment Court clients was not statistically significant, \(\chi^2\)(6, N = 78) = 6.13, p = 0.408.
The difference in recidivism rates between races/ethnicities overall was not statistically significant, \(\chi^2\)(6, N = 551) = 12, p = 0.0619.

Analysis 11: Recidivism Rates by Age

Recidivism rates differed between age groups, but only in Drug Court. Overall, there was a partial step-like decline in recidivism with age. Specifically, under 20-year-olds recidivated at a higher rather than the 50-59-year-olds (note that the sample size for the under 20-year-olds was so small and the margins of error so high that it did not statistically differ from the other age groups). The 20-29-year-olds had significantly higher recidivism rates than all other groups except for the under 20s. The 30-39-year-olds had higher recidivism rates than the 40-49 and 50-59-year olds. All other comparisons were not statistically significant.

There was a statistically significant difference in recidivism rates between age groups among Drug Court clients, \(\chi^2\)(5, N = 384) = 42.63, p < .001.
The following groups in HDC statistically differed:
HDC, <20-yr-olds vs 50-59-yr-olds: \(\chi^2\)(1) = 4.76, p = 0.0291.
HDC, 20-29-yr-olds vs 30-39-yr-olds: \(\chi^2\)(1) = 7.09, p = 0.00774.
HDC, 20-29-yr-olds vs 40-49-yr-olds: \(\chi^2\)(1) = 21.21, p < .001.
HDC, 20-29-yr-olds vs 50-59-yr-olds: \(\chi^2\)(1) = 29.83, p < .001.
HDC, 20-29-yr-olds vs 60-69-yr-olds: \(\chi^2\)(1) = 7.42, p = 0.00644.
HDC, 30-39-yr-olds vs 40-49-yr-olds: \(\chi^2\)(1) = 5.25, p = 0.022.
HDC, 30-39-yr-olds vs 50-59-yr-olds: \(\chi^2\)(1) = 11.94, p < .001.
All other comparisons between groups in HDC were not statistically different.

The difference in recidivism rates between age groups among Mental Health Court clients was not statistically significant, \(\chi^2\)(4, N = 89) = 2.58, p = 0.631.
The difference in recidivism rates between age groups among Veterans Treatment Court clients was not statistically significant, \(\chi^2\)(4, N = 78) = 7.01, p = 0.135.
There was a statistically significant difference in recidivism rates between age groups among all ofenders, \(\chi^2\)(5, N = 551) = 43.71, p < .001.

Graduates Only

Here we focus on recividism among probationers who graduated from the Specialty Courts. This data should be taken on its own. Comparing graduates to the full sample comes with a major risk of drawing incorrect conclusions. Graduates likely differ in many ways - known and unknown - from the full sample. Because of that, it is difficult if not impossible to determine with confidence whether and why recidivism rates differ (or not) between the two groups.

Analysis 12: Overall Recidivism Rates - Graduates

Recidivism rates for graduates were computed for each Specialty Court between CY2014-2021. The estimates for Drug Court graduates are consistent with estimates from a meta-analysis of evaluations of 92 adult drug courts (89 from the United States, two from Canada, and one from Guam). 1

Analysis 13: Recidivism Rates by Yearly Cohort - Graduates

Recidivism rates for each Specialty Court were disaggregated by cohort, in which a cohort was defined as the calendar year in which probationers were admitted to each Specialty Court.

There were no statistically detectable differences in recidivism rates in HDC between 2014 and 2021. There were also no statistical differences in MHC or VTC, and even when all probationers were combined, there were no differences across cohorts.

The difference in recidivism rates between cohorts for Drug Court was not statistically significant, \(\chi^2\)(7, N = 247) = 4.45, p = 0.727.
The difference in recidivism rates between cohorts for Mental Health Court was not statistically significant, \(\chi^2\)(7, N = 43) = 8.37, p = 0.301.
The difference in recidivism rates between cohorts for Veterans Treatment Court was not statistically significant, \(\chi^2\)(6, N = 62) = 10.18, p = 0.117.
The difference in overall recidivism rates between cohorts was not statistically significant, \(\chi^2\)(7, N = 352) = 11.89, p = 0.104.

Analysis 14: Time to Recidivism - Graduates

Recidivism was analyzed by time to recidivate. Longer times to recidivate can indicate gradual desistance from crime. The estimates for Drug Court appear to conform to national estimates of recidivism among drug court graduates. 2

LSI-R and Recidivism

How well did The Level of Service Inventory–Revised (LSI-R) relate to recidivism? Here we look at the predictive and dynamic validity of the LSI-R among the Specialty Court clients.

  • Predictive validity of initial LSI-R
  • Predictive validity of most recent LSI-R
  • Dynamic validity of change from initial to most recent LSI-R

UNDER CONSTRUCTION!

Changes in offense severity among reoffenders

Although some Specialty Court probationers reoffended, a more subtle measure of desistance is evidence that people who reoffend are committing less serious offenses.

UNDER CONSTRUCTION!

Discussion

The results above provide baseline estimates of recidivism rates and time to recidivate for the First Circuit Specialty Courts, including HDC, MHC, and VTC. Generally, recidivism rates did not change over the study period within the Specialty Courts. Although the recidivism rate as defined sat at 52.3%, 43.8%, and 33.3% for HDC, MHC, and VTC, respectively, only 12.2%, 5.6%, and 7.7% of clients were convicted of new charges or had their probation revoked. Time to recidivate averaged between 10.9 and 12.1 months. Property offenders were more likely to recidivate than drug or “other” offenders, but no other differences were statistically detectable. There were no sex/gender differences in recidivism, although there were a few race/ethnicity and age differences.

For graduates of the Specialty Courts, recidivism rate as defined sat at 36.4%, 27.9%, and 22.6% for HDC, MHC, and VTC, respectively. Recidivism rates for graduates did not change over the study period within the Specialty Courts. Only 16.2% of graduates recidivated after 12 months.

Limitations

Comparing to future recidivism rates

If these numbers are used as comparison points for future recidivism studies, they would only be valid if future samples and future circumstances (crime rates, employment rates and other economic factors, capacity for supervision/probation staffing levels, availability of rehabilitation services, etc.) are comparable to the current time frame. More often than not, samples and circumstances change, often in ways that are difficult to quantify and can invalidate comparisons. Given those limitations, these baseline measures may be of limited utility for evaluating the effectiveness of the Specialty Courts over time.

Comparing to historical recidivism rates

Comparisons to past recidivism rates are also fraught. The first fourteen analyses of the current study largely adopted the methods of and applied the recidivism definition from the Interagency Council on Intermediate Sanctions (ICIS) annual statewide recidivism study. However, some important changes were made either to address issues and limitations with prior analyses or to customize analyses for the current study.

Differences with the statewide ICIS recidivism study: Including deferrals

Each supervision case is defined by one and only one charge. The pre-2017 ICIS Recidivism Study identified the highest severity charge on a supervision case and then removed any probationers with deferrals, including deferred no context pleas and deferred guilty pleas. One potential issue with removing deferrals after identifying the highest severity charge is that if a probationer had multiple supervision cases or multiple charges within a case, and only the highest severity charge had a deferral, they would be excluded from the study even though they had other charges not disposed as deferrals.

For the current study, subject matter experts in the Specialty Courts were consulted, and it was decided that deferrals would not be excluded.

Not only does this affect the composition of the sample, it creates a potential difference with the ICIS sample. When the court grants a deferral, it typically does so for first time offenders convicted of less serious offenses. By excluding them from the ICIS recidivism study, the ICIS cohort is potentially on average higher risk than a sample that includes deferrals, although by how much is unknown.

Differences with the statewide ICIS recidivism study: Excluding misdemeanor probationers

Although most of the probationers in the statewide ICIS recidivism study are felony probationers, there are typically also a handful of misdemeanor probationers. Because the Specialty Courts only admit felony probationers, there were no misdemeanor supervision cases in the study sample. Notably, this creates another difference in risk levels between the ICIS recidivism study and the current study. In this case, including misdemeanor probationers would lower the average risk level of the ICIS sample.

Differences with the statewide ICIS recidivism study: The follow-up date

Critically, the ICIS recidivism study defines arrests for the most recent charge after the start of the probation term. In other words, the follow-up date is the probation start date.

However, nearly all of the probationers in the Specialty Courts have already served part of one or more terms of regular probation prior to being admitted into Specialty Court. The follow-up date for Specialty Court probationers is defined not by the start date of the initial probation term but as the date admitted to Specialty Court.

This creates two potentially significant differences with the ICIS recidivism study. The first is that Specialty Court probationers on average have received a greater doasge of programming and services, which could push recidivism rates down. The second is that Specialty Court probationers will on average be older than regular probationers, which could further reduce recidivism rates given the age-crime curve.3 , 4 , 5 Together, these considerations obscure a meaningful, valid comparison between the specialty Courts and the ICIS sample.

Differences with the statewide ICIS recidivism study: Initial risk levels

Higher risk individuals, all else equal, are more likely to recidivate than lower risk probationers. The Specialty Courts are largely composed of people who have failed regular probation at least one time, making up a higher risk sample. This consideration, taken alone, predicts higher recidivism rates among Specialty Court probationers relative to regular probationers. But without a control group, how much this increased risk translates into increased recidivism, and how much of it is blunted by the specialized and intense case planning provided by the Specialty Courts, is unknown.

In practice, this consideration cannot be taken alone but at least has to be considered alongside the treatment and services being provided. Although it cannot be determined with confidence with the data currently available, if Specialty Court reoffense rates were lower than regular probation, it would suggest some impact of the Specialty Courts, although the magnitude of the effect is not straightforward to estimate. It could in fact represent an underestimate of the impact of Specialty Court because the people admitted to these courts necessarily had higher recidivism rates than regular probationers prior to joining the Specialty Courts. In other words, Specialty Court and regular probationers were not at the same starting point with respect to risk to begin with. (Conversely, if the recidivism rates for Specialty Court were higher than regular probation, it would not necessarily imply the Specialty Courts were ineffective). With this type of secondary data and retrospective observational design, there are many moving parts that make it difficult to know with confidence the effectiveness of the Specialty Courts.

Comparing between Specialty Courts

Given the limits of observational data, attempts to compare recidivism estimates between each of the Specialty Courts comes with the risk of drawing invalid conclusions, or at least conclusions that rest on weak, ambiguous evidence. For instance, although there are apparent differences in the recidivism rates between the Specialty Courts, comparing between them is not straightforward for a number of reasons. For example, the average age of clients in the three programs vary from 35.8 years in HDC, 40.2 years in MHC, and 42.6 years in VTC. Given the age-crime curve, a difference in recidivism rates is expected simply due to the age differences across the three groups.

Future directions

The current approach to studying the Specialty Courts relies on secondary data and a non-experimental research design. The recidivism estimates from this type of research provides some descriptive data, but due to the various uncontrolled differences between groups and other factors not being held constant, the evidence available is generally insufficient to answer questions about relative effectiveness, absolute effectiveness, and whether effectiveness of the Specialty Courts has changed over time.

An alternative approach to studying the Specialty Courts that goes beyond the limitations of an observational research design would include running randomized control trials (RCTs) to study the effectiveness of various interventions and programs related to rehabilitation. When RCTs/experiments are not feasible, another approach would involve using causal inference methods with quasi-experimental designs. To use causal inference, a strong theory about the causes of recidivism would be necessary, as well as careful measurement on those causal factors to determine how much each one contributes to recidivism. Both RCTs and causal inference approaches could assist in identifying the primary root causes of reoffending and what aspects of the Specialty Courts are and are not effective at reducing recidivism.

Acknowledgments

Special thanks to Greg Uwono, Specialty Court Administrator, and Specialty Court Judges Clarissa Malinao and Matthew Viola, for working closely with the research team in developing the recidivism study. Special recognition goes to Greg for updating and checking data discrepancies and missing data, and answering many, many questions about the Specialty Courts. Mahalo nui to Judges Malinao and Viola for allowing us to observe status review hearings and answering questions.

Research assistance was provided by Ian Strieter, PhD student in Psychology at UH Manoa, and Jourdan Owens, MS student in Counseling Psychology at Chaminade University. Finally, a big mahalo to Vincent Borja, ICIS Project Coordinator, for providing LSI-R data and to Wendy Char and her team at the Hawaii Criminal Justice Data Center within the Office of the Attorney General for pulling rearrest records for this study.

Footnotes

  1. Mitchell, O., Wilson, D. B., Eggers, A., & MacKenzie, D. L. (2012). Assessing the effectiveness of drug courts on recidivism: A meta-analytic review of traditional and non-traditional drug courts. Journal of Criminal Justice, 40, 60-71..↩︎

  2. Roman, J., Townsend, W., & Bhati, A. S. (2003). Recidivism Rates for Drug Court Graduates: Nationally based estimates. Urban Institute..↩︎

  3. Farrington, D. P. (1986). Age and crime. Crime and justice, 7, 189-250.↩︎

  4. Hirschi, T., & Gottfredson, M. (1983). Age and the explanation of crime. American journal of sociology, 89, 552-584.↩︎

  5. Wilson, M., & Daly, M. (1985). Competitiveness, risk taking, and violence: The young male syndrome. Ethology and sociobiology, 6, 59-73.↩︎